OverviewVersionsSchemas
v1.0.0·3,520 records·0 files·2.4 MB
Jul 9, 20261
Author
2,716 records
Presentation
754 records
Session
6 records
TopicArea
44 records
README

JoCN Forum — CNS Annual Meeting Abstracts

Poster abstracts from the Cognitive Neuroscience Society (CNS) Annual Meeting program, harvested for the Journal of Cognitive Neuroscience Forum (JoCNForum).

JoCN currently posts these abstracts to their PubPub forum by hand — trainees open each poster, copy the title/authors/body, create a pub, and mint a DOI. This collection automates the read half of that workflow and, critically, captures the topic-area taxonomy the manual process drops on the floor. CNS tags every presentation with a hierarchical CATEGORY: Subcategory topic (e.g. ATTENTION: Auditory), which is exactly the thematic label the forum can't organize by today.

Scraped from public, server-rendered HTML (no auth):

  • Session index — https://www.cogneurosociety.org/poster-sessions/
  • Session listing — https://www.cogneurosociety.org/poster-sessions/?id=<sessionId>
  • Abstract page — https://www.cogneurosociety.org/poster/?id=<posterId>

Record types

Type Description
Presentation A poster with title, full abstract, author list, per-author affiliations, corresponding email, session, and topic area. id is the CNS poster id (stable primary key).
Author Presenting + co-authors, deduplicated by normalized full name, with a count of presentations.
Session The six poster sessions (A–F), with day, room, and schedule.
TopicArea The CNS topic taxonomy (CATEGORY: Subcategory), split into category / subcategory, with per-topic presentation counts. This is the field that lets abstracts be grouped thematically the way they appear in the program.

Relationships are expressed by reference: a Presentation stores sessionId, topicArea (a TopicArea id), and authorIds; it also carries an authorships array giving each author's institution(s) for that specific presentation.

Scale

~754 presentations across 6 sessions, ~2,700 unique authors, 44 topic areas (10 top-level categories). Full abstracts are present for every presentation. About 17% of posters do not expose a corresponding email in the source — that field is null for those (it is not a scrape gap).

Scraping

pnpm scrape jocn/forum
pnpm push jocn/forum --env dev

Every fetched page is cached under data/.cache/ (gitignored), so the crawl is resumable and reruns don't re-hit the CNS site. Requests are throttled (bounded concurrency + delay, exponential backoff on 429/5xx). SCRAPE_MAX_POSTERS=N caps the crawl for a quick validation run.

Not captured

  • DOIs — CNS abstract pages carry no DOI; JoCN mints those in PubPub. The sourceUrl links back to the canonical CNS abstract page.
  • Symposia / Data Blitz / Rising Stars — this collection covers poster sessions only (the bulk of the program). The same two-level scrape applies to the other presentation types if they're brought into scope later.